Very Important Statistical Analyses

Hold on to your hats and glasses folks, because this here is the boringest ride on the internet. I'm going to present to you some shoddy statistical analysis I've done that you can use to impress your friends and colleauges.

First, I wanted to point out that most bloggers are quitters. Stupid quitters. Well over half of all bloggers (at least those that show up in the profile viewer on blogger.com) have made less than 10 posts to their blogs. The overall average is about 38 posts, but that is highly skewed by the few that have made hundreds (yours truly is not quite in that rarified atmosphere yet. I'm at 163). The median number is actually 2. Here's a chart that shows the frequency distribution.




Now there are a couple things to note here. First, many of the people in the sample may not actually have blogspot based blogs and only have a username in order to comment. This would skew the data lower. Second, there are certain demographics of people that are much more likely to maintain their blogs. For instance, fans of The Shins, and The Postal Service (it says so in their profiles) are much more likely to keep up with posting than are fans of say, The Rolling Stones, though I have not quantified the relative differences. I think it's a generational thing.


On to the second analysis that I have been working on in fits and starts for quite a while now: How many pseudo-random clicks does it take to uncover a bomb in MS Minesweeper? In the chart a result of "2" means that I managed 2 clicks without uncovering a bomb, i.e. the bomb was found on the third click. Here are the results of 200 trials...



First, please note that there were no games in which I uncovered a bomb with the first click. I believe this is programmed into the game. The result of this is that I am slightly more likely than normal to find one on my second click, which the results tend to show. However, due to the presence of the non-bomb programming, my average number of clicks was slightly higher than would be expected if the program was totally random (4.98 vs 4.85). Also, take these results with a grain of salt becuase my clicks were not actually random, they were pseudo-random in that I was intentionally forcing an appearance of randomness by clicking all over the screen and very rarely in close proximity to the previous click.

A Ton of Bricks: doing pointless and ridiculous statistics so you don't have to.